Establishment of Molecular Classification Models for Early Diagnosis of Digestive System Cancers

Design of the project: 1) Participants who sign informed consent forms will complete the disease history information collection;2) Collect samples of cancer tissue and corresponding paracancer samples, as well as peripheral blood; 3) Using the MONOD patent detection data of the company's previous research and combining with literature retrieval, to analyze the tissue samples of digestive system cancers to screen the methylation mutation sites;4) ctDNA methylation markers are preliminarily screened by plasma of healthy persons paired with tissues;5) Using the screened blood ctDNA specific methylation markers together with other molecular markers to establish a targeted detection method for early digestive system cancer lesions;6) Retain the most distinguishing targets in the original panel test and establish the second version of the panel;7) Establish and validate predictive models to finalize multitype molecular target detection for peripheral blood DNA;8) Construct statistical learning models based on ctDNA methylation and fragment features, miRNA7™ and CTC, and optimize the models respectively;9) Evaluate the sensitivity and specificity of the models in cancer early detection.